Transfer Learning in Keras

The Jupyter notebook described in the video can be accessed from the aind2-cnn GitHub repository linked here. Navigate to the transfer-learning/ folder and open transfer_learning.ipynb. If you'd like to learn how to calculate your own bottleneck features, look at bottleneck_features.ipynb. (You may have trouble running bottleneck_features.ipynb on an AWS GPU instance - if so, feel free to use the notebook on your local CPU/GPU instead!)

Optional Resources

Here's the first research paper to propose GAP layers for object localization.
Check out this repository that uses a CNN for object localization.
Watch this video demonstration of object localization with a CNN.
Check out this repository that uses visualization techniques to better understand bottleneck features.